Titles
stringlengths 6
220
| Abstracts
stringlengths 37
3.26k
| Years
int64 1.99k
2.02k
| Categories
stringclasses 1
value |
|---|---|---|---|
Semi-supervised Interactive Intent Labeling
|
Building the Natural Language Understanding (NLU) modules of task-oriented
Spoken Dialogue Systems (SDS) involves a definition of intents and entities,
collection of task-relevant data, annotating the data with intents and
entities, and then repeating the same process over and over again for adding
any functionality/enhancement to the SDS. In this work, we showcase an Intent
Bulk Labeling system where SDS developers can interactively label and augment
training data from unlabeled utterance corpora using advanced clustering and
visual labeling methods. We extend the Deep Aligned Clustering work with a
better backbone BERT model, explore techniques to select the seed data for
labeling, and develop a data balancing method using an oversampling technique
that utilizes paraphrasing models. We also look at the effect of data
augmentation on the clustering process. Our results show that we can achieve
over 10% gain in clustering accuracy on some datasets using the combination of
the above techniques. Finally, we extract utterance embeddings from the
clustering model and plot the data to interactively bulk label the samples,
reducing the time and effort for data labeling of the whole dataset
significantly.
| 2,021 |
Computation and Language
|
Named Entity Recognition and Linking Augmented with Large-Scale
Structured Data
|
In this paper we describe our submissions to the 2nd and 3rd SlavNER Shared
Tasks held at BSNLP 2019 and BSNLP 2021, respectively. The tasks focused on the
analysis of Named Entities in multilingual Web documents in Slavic languages
with rich inflection. Our solution takes advantage of large collections of both
unstructured and structured documents. The former serve as data for
unsupervised training of language models and embeddings of lexical units. The
latter refers to Wikipedia and its structured counterpart - Wikidata, our
source of lemmatization rules, and real-world entities. With the aid of those
resources, our system could recognize, normalize and link entities, while being
trained with only small amounts of labeled data.
| 2,021 |
Computation and Language
|
Towards Clinical Encounter Summarization: Learning to Compose Discharge
Summaries from Prior Notes
|
The records of a clinical encounter can be extensive and complex, thus
placing a premium on tools that can extract and summarize relevant information.
This paper introduces the task of generating discharge summaries for a clinical
encounter. Summaries in this setting need to be faithful, traceable, and scale
to multiple long documents, motivating the use of extract-then-abstract
summarization cascades. We introduce two new measures, faithfulness and
hallucination rate for evaluation in this task, which complement existing
measures for fluency and informativeness. Results across seven medical sections
and five models show that a summarization architecture that supports
traceability yields promising results, and that a sentence-rewriting approach
performs consistently on the measure used for faithfulness
(faithfulness-adjusted $F_3$) over a diverse range of generated sections.
| 2,021 |
Computation and Language
|
AraStance: A Multi-Country and Multi-Domain Dataset of Arabic Stance
Detection for Fact Checking
|
With the continuing spread of misinformation and disinformation online, it is
of increasing importance to develop combating mechanisms at scale in the form
of automated systems that support multiple languages. One task of interest is
claim veracity prediction, which can be addressed using stance detection with
respect to relevant documents retrieved online. To this end, we present our new
Arabic Stance Detection dataset (AraStance) of 4,063 claim--article pairs from
a diverse set of sources comprising three fact-checking websites and one news
website. AraStance covers false and true claims from multiple domains (e.g.,
politics, sports, health) and several Arab countries, and it is well-balanced
between related and unrelated documents with respect to the claims. We
benchmark AraStance, along with two other stance detection datasets, using a
number of BERT-based models. Our best model achieves an accuracy of 85\% and a
macro F1 score of 78\%, which leaves room for improvement and reflects the
challenging nature of AraStance and the task of stance detection in general.
| 2,021 |
Computation and Language
|
Multi-view Inference for Relation Extraction with Uncertain Knowledge
|
Knowledge graphs (KGs) are widely used to facilitate relation extraction (RE)
tasks. While most previous RE methods focus on leveraging deterministic KGs,
uncertain KGs, which assign a confidence score for each relation instance, can
provide prior probability distributions of relational facts as valuable
external knowledge for RE models. This paper proposes to exploit uncertain
knowledge to improve relation extraction. Specifically, we introduce ProBase,
an uncertain KG that indicates to what extent a target entity belongs to a
concept, into our RE architecture. We then design a novel multi-view inference
framework to systematically integrate local context and global knowledge across
three views: mention-, entity- and concept-view. The experimental results show
that our model achieves competitive performances on both sentence- and
document-level relation extraction, which verifies the effectiveness of
introducing uncertain knowledge and the multi-view inference framework that we
design.
| 2,021 |
Computation and Language
|
MelBERT: Metaphor Detection via Contextualized Late Interaction using
Metaphorical Identification Theories
|
Automated metaphor detection is a challenging task to identify metaphorical
expressions of words in a sentence. To tackle this problem, we adopt
pre-trained contextualized models, e.g., BERT and RoBERTa. To this end, we
propose a novel metaphor detection model, namely metaphor-aware late
interaction over BERT (MelBERT). Our model not only leverages contextualized
word representation but also benefits from linguistic metaphor identification
theories to distinguish between the contextual and literal meaning of words.
Our empirical results demonstrate that MelBERT outperforms several strong
baselines on four benchmark datasets, i.e., VUA-18, VUA-20, MOH-X, and TroFi.
| 2,021 |
Computation and Language
|
SELF & FEIL: Emotion and Intensity Lexicons for Finnish
|
This paper introduces a Sentiment and Emotion Lexicon for Finnish (SELF) and
a Finnish Emotion Intensity Lexicon (FEIL). We describe the lexicon creation
process and evaluate the lexicon using some commonly available tools. The
lexicon uses annotations projected from the NRC Emotion Lexicon with carefully
edited translations. To our knowledge, this is the first comprehensive
sentiment and emotion lexicon for Finnish.
| 2,021 |
Computation and Language
|
PCFGs Can Do Better: Inducing Probabilistic Context-Free Grammars with
Many Symbols
|
Probabilistic context-free grammars (PCFGs) with neural parameterization have
been shown to be effective in unsupervised phrase-structure grammar induction.
However, due to the cubic computational complexity of PCFG representation and
parsing, previous approaches cannot scale up to a relatively large number of
(nonterminal and preterminal) symbols. In this work, we present a new
parameterization form of PCFGs based on tensor decomposition, which has at most
quadratic computational complexity in the symbol number and therefore allows us
to use a much larger number of symbols. We further use neural parameterization
for the new form to improve unsupervised parsing performance. We evaluate our
model across ten languages and empirically demonstrate the effectiveness of
using more symbols. Our code: https://github.com/sustcsonglin/TN-PCFG
| 2,021 |
Computation and Language
|
Gradient-based Adversarial Attacks against Text Transformers
|
We propose the first general-purpose gradient-based attack against
transformer models. Instead of searching for a single adversarial example, we
search for a distribution of adversarial examples parameterized by a
continuous-valued matrix, hence enabling gradient-based optimization. We
empirically demonstrate that our white-box attack attains state-of-the-art
attack performance on a variety of natural language tasks. Furthermore, we show
that a powerful black-box transfer attack, enabled by sampling from the
adversarial distribution, matches or exceeds existing methods, while only
requiring hard-label outputs.
| 2,021 |
Computation and Language
|
Evaluating Document Representations for Content-based Legal Literature
Recommendations
|
Recommender systems assist legal professionals in finding relevant literature
for supporting their case. Despite its importance for the profession, legal
applications do not reflect the latest advances in recommender systems and
representation learning research. Simultaneously, legal recommender systems are
typically evaluated in small-scale user study without any public available
benchmark datasets. Thus, these studies have limited reproducibility. To
address the gap between research and practice, we explore a set of
state-of-the-art document representation methods for the task of retrieving
semantically related US case law. We evaluate text-based (e.g., fastText,
Transformers), citation-based (e.g., DeepWalk, Poincar\'e), and hybrid methods.
We compare in total 27 methods using two silver standards with annotations for
2,964 documents. The silver standards are newly created from Open Case Book and
Wikisource and can be reused under an open license facilitating
reproducibility. Our experiments show that document representations from
averaged fastText word vectors (trained on legal corpora) yield the best
results, closely followed by Poincar\'e citation embeddings. Combining fastText
and Poincar\'e in a hybrid manner further improves the overall result. Besides
the overall performance, we analyze the methods depending on document length,
citation count, and the coverage of their recommendations. We make our source
code, models, and datasets publicly available at
https://github.com/malteos/legal-document-similarity/.
| 2,021 |
Computation and Language
|
Removing Word-Level Spurious Alignment between Images and
Pseudo-Captions in Unsupervised Image Captioning
|
Unsupervised image captioning is a challenging task that aims at generating
captions without the supervision of image-sentence pairs, but only with images
and sentences drawn from different sources and object labels detected from the
images. In previous work, pseudo-captions, i.e., sentences that contain the
detected object labels, were assigned to a given image. The focus of the
previous work was on the alignment of input images and pseudo-captions at the
sentence level. However, pseudo-captions contain many words that are irrelevant
to a given image. In this work, we investigate the effect of removing
mismatched words from image-sentence alignment to determine how they make this
task difficult. We propose a simple gating mechanism that is trained to align
image features with only the most reliable words in pseudo-captions: the
detected object labels. The experimental results show that our proposed method
outperforms the previous methods without introducing complex sentence-level
learning objectives. Combined with the sentence-level alignment method of
previous work, our method further improves its performance. These results
confirm the importance of careful alignment in word-level details.
| 2,021 |
Computation and Language
|
Improving BERT Model Using Contrastive Learning for Biomedical Relation
Extraction
|
Contrastive learning has been used to learn a high-quality representation of
the image in computer vision. However, contrastive learning is not widely
utilized in natural language processing due to the lack of a general method of
data augmentation for text data. In this work, we explore the method of
employing contrastive learning to improve the text representation from the BERT
model for relation extraction. The key knob of our framework is a unique
contrastive pre-training step tailored for the relation extraction tasks by
seamlessly integrating linguistic knowledge into the data augmentation.
Furthermore, we investigate how large-scale data constructed from the external
knowledge bases can enhance the generality of contrastive pre-training of BERT.
The experimental results on three relation extraction benchmark datasets
demonstrate that our method can improve the BERT model representation and
achieve state-of-the-art performance. In addition, we explore the
interpretability of models by showing that BERT with contrastive pre-training
relies more on rationales for prediction. Our code and data are publicly
available at: https://github.com/udel-biotm-lab/BERT-CLRE.
| 2,021 |
Computation and Language
|
Learning Syntax from Naturally-Occurring Bracketings
|
Naturally-occurring bracketings, such as answer fragments to natural language
questions and hyperlinks on webpages, can reflect human syntactic intuition
regarding phrasal boundaries. Their availability and approximate correspondence
to syntax make them appealing as distant information sources to incorporate
into unsupervised constituency parsing. But they are noisy and incomplete; to
address this challenge, we develop a partial-brackets-aware structured ramp
loss in learning. Experiments demonstrate that our distantly-supervised models
trained on naturally-occurring bracketing data are more accurate in inducing
syntactic structures than competing unsupervised systems. On the English WSJ
corpus, our models achieve an unlabeled F1 score of 68.9 for constituency
parsing.
| 2,021 |
Computation and Language
|
Diversity-Aware Batch Active Learning for Dependency Parsing
|
While the predictive performance of modern statistical dependency parsers
relies heavily on the availability of expensive expert-annotated treebank data,
not all annotations contribute equally to the training of the parsers. In this
paper, we attempt to reduce the number of labeled examples needed to train a
strong dependency parser using batch active learning (AL). In particular, we
investigate whether enforcing diversity in the sampled batches, using
determinantal point processes (DPPs), can improve over their diversity-agnostic
counterparts. Simulation experiments on an English newswire corpus show that
selecting diverse batches with DPPs is superior to strong selection strategies
that do not enforce batch diversity, especially during the initial stages of
the learning process. Additionally, our diversityaware strategy is robust under
a corpus duplication setting, where diversity-agnostic sampling strategies
exhibit significant degradation.
| 2,021 |
Computation and Language
|
RECKONition: a NLP-based system for Industrial Accidents at Work
Prevention
|
Extracting patterns and useful information from Natural Language datasets is
a challenging task, especially when dealing with data written in a language
different from English, like Italian. Machine and Deep Learning, together with
Natural Language Processing (NLP) techniques have widely spread and improved
lately, providing a plethora of useful methods to address both Supervised and
Unsupervised problems on textual information. We propose RECKONition, a
NLP-based system for Industrial Accidents at Work Prevention. RECKONition,
which is meant to provide Natural Language Understanding, Clustering and
Inference, is the result of a joint partnership with the Italian National
Institute for Insurance against Accidents at Work (INAIL). The obtained results
showed the ability to process textual data written in Italian describing
industrial accidents dynamics and consequences.
| 2,021 |
Computation and Language
|
Variable-Length Codes Independent or Closed with respect to Edit
Relations
|
We investigate inference of variable-length codes in other domains of
computer science, such as noisy information transmission or information
retrieval-storage: in such topics, traditionally mostly constant-length
codewords act. The study is relied upon the two concepts of independent and
closed sets. We focus to those word relations whose images are computed by
applying some peculiar combinations of deletion, insertion, or substitution. In
particular, characterizations of variable-length codes that are maximal in the
families of $\tau$-independent or $\tau$-closed codes are provided.
| 2,021 |
Computation and Language
|
Using Fisher's Exact Test to Evaluate Association Measures for N-grams
|
To determine whether some often-used lexical association measures assign high
scores to n-grams that chance could have produced as frequently as observed, we
used an extension of Fisher's exact test to sequences longer than two words to
analyse a corpus of four million words. The results, based on the
precision-recall curve and a new index called chance-corrected average
precision, show that, as expected, simple-ll is extremely effective. They also
show, however, that MI3 is more efficient than the other hypothesis tests-based
measures and even reaches a performance level almost equal to simple-ll for
3-grams. It is additionally observed that some measures are more efficient for
3-grams than for 2-grams, while others stagnate.
| 2,021 |
Computation and Language
|
How (Non-)Optimal is the Lexicon?
|
The mapping of lexical meanings to wordforms is a major feature of natural
languages. While usage pressures might assign short words to frequent meanings
(Zipf's law of abbreviation), the need for a productive and open-ended
vocabulary, local constraints on sequences of symbols, and various other
factors all shape the lexicons of the world's languages. Despite their
importance in shaping lexical structure, the relative contributions of these
factors have not been fully quantified. Taking a coding-theoretic view of the
lexicon and making use of a novel generative statistical model, we define upper
bounds for the compressibility of the lexicon under various constraints.
Examining corpora from 7 typologically diverse languages, we use those upper
bounds to quantify the lexicon's optimality and to explore the relative costs
of major constraints on natural codes. We find that (compositional) morphology
and graphotactics can sufficiently account for most of the complexity of
natural codes -- as measured by code length.
| 2,021 |
Computation and Language
|
MOROCCO: Model Resource Comparison Framework
|
The new generation of pre-trained NLP models push the SOTA to the new limits,
but at the cost of computational resources, to the point that their use in real
production environments is often prohibitively expensive. We tackle this
problem by evaluating not only the standard quality metrics on downstream tasks
but also the memory footprint and inference time. We present MOROCCO, a
framework to compare language models compatible with \texttt{jiant} environment
which supports over 50 NLU tasks, including SuperGLUE benchmark and multiple
probing suites. We demonstrate its applicability for two GLUE-like suites in
different languages.
| 2,021 |
Computation and Language
|
Dynabench: Rethinking Benchmarking in NLP
|
We introduce Dynabench, an open-source platform for dynamic dataset creation
and model benchmarking. Dynabench runs in a web browser and supports
human-and-model-in-the-loop dataset creation: annotators seek to create
examples that a target model will misclassify, but that another person will
not. In this paper, we argue that Dynabench addresses a critical need in our
community: contemporary models quickly achieve outstanding performance on
benchmark tasks but nonetheless fail on simple challenge examples and falter in
real-world scenarios. With Dynabench, dataset creation, model development, and
model assessment can directly inform each other, leading to more robust and
informative benchmarks. We report on four initial NLP tasks, illustrating these
concepts and highlighting the promise of the platform, and address potential
objections to dynamic benchmarking as a new standard for the field.
| 2,021 |
Computation and Language
|
Bridging the gap between streaming and non-streaming ASR systems
bydistilling ensembles of CTC and RNN-T models
|
Streaming end-to-end automatic speech recognition (ASR) systems are widely
used in everyday applications that require transcribing speech to text in
real-time. Their minimal latency makes them suitable for such tasks. Unlike
their non-streaming counterparts, streaming models are constrained to be causal
with no future context and suffer from higher word error rates (WER). To
improve streaming models, a recent study [1] proposed to distill a
non-streaming teacher model on unsupervised utterances, and then train a
streaming student using the teachers' predictions. However, the performance gap
between teacher and student WERs remains high. In this paper, we aim to close
this gap by using a diversified set of non-streaming teacher models and
combining them using Recognizer Output Voting Error Reduction (ROVER). In
particular, we show that, despite being weaker than RNN-T models, CTC models
are remarkable teachers. Further, by fusing RNN-T and CTC models together, we
build the strongest teachers. The resulting student models drastically improve
upon streaming models of previous work [1]: the WER decreases by 41% on
Spanish, 27% on Portuguese, and 13% on French.
| 2,021 |
Computation and Language
|
Impact of Encoding and Segmentation Strategies on End-to-End
Simultaneous Speech Translation
|
Boosted by the simultaneous translation shared task at IWSLT 2020, promising
end-to-end online speech translation approaches were recently proposed. They
consist in incrementally encoding a speech input (in a source language) and
decoding the corresponding text (in a target language) with the best possible
trade-off between latency and translation quality. This paper investigates two
key aspects of end-to-end simultaneous speech translation: (a) how to encode
efficiently the continuous speech flow, and (b) how to segment the speech flow
in order to alternate optimally between reading (R: encoding input) and writing
(W: decoding output) operations. We extend our previously proposed end-to-end
online decoding strategy and show that while replacing BLSTM by ULSTM encoding
degrades performance in offline mode, it actually improves both efficiency and
performance in online mode. We also measure the impact of different methods to
segment the speech signal (using fixed interval boundaries, oracle word
boundaries or randomly set boundaries) and show that our best end-to-end online
decoding strategy is surprisingly the one that alternates R/W operations on
fixed size blocks on our English-German speech translation setup.
| 2,021 |
Computation and Language
|
Experts, Errors, and Context: A Large-Scale Study of Human Evaluation
for Machine Translation
|
Human evaluation of modern high-quality machine translation systems is a
difficult problem, and there is increasing evidence that inadequate evaluation
procedures can lead to erroneous conclusions. While there has been considerable
research on human evaluation, the field still lacks a commonly-accepted
standard procedure. As a step toward this goal, we propose an evaluation
methodology grounded in explicit error analysis, based on the Multidimensional
Quality Metrics (MQM) framework. We carry out the largest MQM research study to
date, scoring the outputs of top systems from the WMT 2020 shared task in two
language pairs using annotations provided by professional translators with
access to full document context. We analyze the resulting data extensively,
finding among other results a substantially different ranking of evaluated
systems from the one established by the WMT crowd workers, exhibiting a clear
preference for human over machine output. Surprisingly, we also find that
automatic metrics based on pre-trained embeddings can outperform human crowd
workers. We make our corpus publicly available for further research.
| 2,022 |
Computation and Language
|
AMR Parsing with Action-Pointer Transformer
|
Abstract Meaning Representation parsing is a sentence-to-graph prediction
task where target nodes are not explicitly aligned to sentence tokens. However,
since graph nodes are semantically based on one or more sentence tokens,
implicit alignments can be derived. Transition-based parsers operate over the
sentence from left to right, capturing this inductive bias via alignments at
the cost of limited expressiveness. In this work, we propose a transition-based
system that combines hard-attention over sentences with a target-side action
pointer mechanism to decouple source tokens from node representations and
address alignments. We model the transitions as well as the pointer mechanism
through straightforward modifications within a single Transformer architecture.
Parser state and graph structure information are efficiently encoded using
attention heads. We show that our action-pointer approach leads to increased
expressiveness and attains large gains (+1.6 points) against the best
transition-based AMR parser in very similar conditions. While using no graph
re-categorization, our single model yields the second best Smatch score on AMR
2.0 (81.8), which is further improved to 83.4 with silver data and ensemble
decoding.
| 2,021 |
Computation and Language
|
Entailment as Few-Shot Learner
|
Large pre-trained language models (LMs) have demonstrated remarkable ability
as few-shot learners. However, their success hinges largely on scaling model
parameters to a degree that makes it challenging to train and serve. In this
paper, we propose a new approach, named as EFL, that can turn small LMs into
better few-shot learners. The key idea of this approach is to reformulate
potential NLP task into an entailment one, and then fine-tune the model with as
little as 8 examples. We further demonstrate our proposed method can be: (i)
naturally combined with an unsupervised contrastive learning-based data
augmentation method; (ii) easily extended to multilingual few-shot learning. A
systematic evaluation on 18 standard NLP tasks demonstrates that this approach
improves the various existing SOTA few-shot learning methods by 12\%, and
yields competitive few-shot performance with 500 times larger models, such as
GPT-3.
| 2,021 |
Computation and Language
|
Let's Play Mono-Poly: BERT Can Reveal Words' Polysemy Level and
Partitionability into Senses
|
Pre-trained language models (LMs) encode rich information about linguistic
structure but their knowledge about lexical polysemy remains unclear. We
propose a novel experimental setup for analysing this knowledge in LMs
specifically trained for different languages (English, French, Spanish and
Greek) and in multilingual BERT. We perform our analysis on datasets carefully
designed to reflect different sense distributions, and control for parameters
that are highly correlated with polysemy such as frequency and grammatical
category. We demonstrate that BERT-derived representations reflect words'
polysemy level and their partitionability into senses. Polysemy-related
information is more clearly present in English BERT embeddings, but models in
other languages also manage to establish relevant distinctions between words at
different polysemy levels. Our results contribute to a better understanding of
the knowledge encoded in contextualised representations and open up new avenues
for multilingual lexical semantics research.
| 2,021 |
Computation and Language
|
The Zero Resource Speech Challenge 2021: Spoken language modelling
|
We present the Zero Resource Speech Challenge 2021, which asks participants
to learn a language model directly from audio, without any text or labels. The
challenge is based on the Libri-light dataset, which provides up to 60k hours
of audio from English audio books without any associated text. We provide a
pipeline baseline system consisting on an encoder based on contrastive
predictive coding (CPC), a quantizer ($k$-means) and a standard language model
(BERT or LSTM). The metrics evaluate the learned representations at the
acoustic (ABX discrimination), lexical (spot-the-word), syntactic
(acceptability judgment) and semantic levels (similarity judgment). We present
an overview of the eight submitted systems from four groups and discuss the
main results.
| 2,021 |
Computation and Language
|
Adapting Coreference Resolution for Processing Violent Death Narratives
|
Coreference resolution is an important component in analyzing narrative text
from administrative data (e.g., clinical or police sources). However, existing
coreference models trained on general language corpora suffer from poor
transferability due to domain gaps, especially when they are applied to
gender-inclusive data with lesbian, gay, bisexual, and transgender (LGBT)
individuals. In this paper, we analyzed the challenges of coreference
resolution in an exemplary form of administrative text written in English:
violent death narratives from the USA's Centers for Disease Control's (CDC)
National Violent Death Reporting System. We developed a set of data
augmentation rules to improve model performance using a probabilistic data
programming framework. Experiments on narratives from an administrative
database, as well as existing gender-inclusive coreference datasets,
demonstrate the effectiveness of data augmentation in training coreference
models that can better handle text data about LGBT individuals.
| 2,021 |
Computation and Language
|
Cross-lingual hate speech detection based on multilingual
domain-specific word embeddings
|
Automatic hate speech detection in online social networks is an important
open problem in Natural Language Processing (NLP). Hate speech is a
multidimensional issue, strongly dependant on language and cultural factors.
Despite its relevance, research on this topic has been almost exclusively
devoted to English. Most supervised learning resources, such as labeled
datasets and NLP tools, have been created for this same language. Considering
that a large portion of users worldwide speak in languages other than English,
there is an important need for creating efficient approaches for multilingual
hate speech detection. In this work we propose to address the problem of
multilingual hate speech detection from the perspective of transfer learning.
Our goal is to determine if knowledge from one particular language can be used
to classify other language, and to determine effective ways to achieve this. We
propose a hate specific data representation and evaluate its effectiveness
against general-purpose universal representations most of which, unlike our
proposed model, have been trained on massive amounts of data. We focus on a
cross-lingual setting, in which one needs to classify hate speech in one
language without having access to any labeled data for that language. We show
that the use of our simple yet specific multilingual hate representations
improves classification results. We explain this with a qualitative analysis
showing that our specific representation is able to capture some common
patterns in how hate speech presents itself in different languages.
Our proposal constitutes, to the best of our knowledge, the first attempt for
constructing multilingual specific-task representations. Despite its
simplicity, our model outperformed the previous approaches for most of the
experimental setups. Our findings can orient future solutions toward the use of
domain-specific representations.
| 2,021 |
Computation and Language
|
A Survey on sentiment analysis in Persian: A Comprehensive System
Perspective Covering Challenges and Advances in Resources, and Methods
|
Social media has been remarkably grown during the past few years. Nowadays,
posting messages on social media websites has become one of the most popular
Internet activities. The vast amount of user-generated content has made social
media the most extensive data source of public opinion. Sentiment analysis is
one of the techniques used to analyze user-generated data. The Persian language
has specific features and thereby requires unique methods and models to be
adopted for sentiment analysis, which are different from those in English
language. Sentiment analysis in each language has specified prerequisites;
hence, the direct use of methods, tools, and resources developed for English
language in Persian has its limitations. The main target of this paper is to
provide a comprehensive literature survey for state-of-the-art advances in
Persian sentiment analysis. In this regard, the present study aims to
investigate and compare the previous sentiment analysis studies on Persian
texts and describe contributions presented in articles published in the last
decade. First, the levels, approaches, and tasks for sentiment analysis are
described. Then, a detailed survey of the sentiment analysis methods used for
Persian texts is presented, and previous relevant works on Persian Language are
discussed. Moreover, we present in this survey the authentic and published
standard sentiment analysis resources and advances that have been done for
Persian sentiment analysis. Finally, according to the state-of-the-art
development of English sentiment analysis, some issues and challenges not being
addressed in Persian texts are listed, and some guidelines and trends are
provided for future research on Persian texts. The paper provides information
to help new or established researchers in the field as well as industry
developers who aim to deploy an operational complete sentiment analysis system.
| 2,021 |
Computation and Language
|
An Adversarial Transfer Network for Knowledge Representation Learning
|
Knowledge representation learning has received a lot of attention in the past
few years. The success of existing methods heavily relies on the quality of
knowledge graphs. The entities with few triplets tend to be learned with less
expressive power. Fortunately, there are many knowledge graphs constructed from
various sources, the representations of which could contain much information.
We propose an adversarial embedding transfer network ATransN, which transfers
knowledge from one or more teacher knowledge graphs to a target one through an
aligned entity set without explicit data leakage. Specifically, we add soft
constraints on aligned entity pairs and neighbours to the existing knowledge
representation learning methods. To handle the problem of possible distribution
differences between teacher and target knowledge graphs, we introduce an
adversarial adaption module. The discriminator of this module evaluates the
degree of consistency between the embeddings of an aligned entity pair. The
consistency score is then used as the weights of soft constraints. It is not
necessary to acquire the relations and triplets in teacher knowledge graphs
because we only utilize the entity representations. Knowledge graph completion
results show that ATransN achieves better performance against baselines without
transfer on three datasets, CN3l, WK3l, and DWY100k. The ablation study
demonstrates that ATransN can bring steady and consistent improvement in
different settings. The extension of combining other knowledge graph embedding
algorithms and the extension with three teacher graphs display the promising
generalization of the adversarial transfer network.
| 2,021 |
Computation and Language
|
Event-driven timeseries analysis and the comparison of public reactions
on COVID-19
|
The rapid spread of COVID-19 has already affected human lives throughout the
globe. Governments of different countries have taken various measures, but how
they affected people lives is not clear. In this study, a rule-based and a
machine-learning based models are applied to answer the above question using
public tweets from Japan, USA, UK, and Australia. Two polarity timeseries
(meanPol and pnRatio) and two events, namely "lockdown or emergency (LED)" and
"the economic support package (ESP)", are considered in this study. Statistical
testing on the sub-series around LED and ESP events showed their positive
impacts to the people of (UK and Australia) and (USA and UK), respectively
unlike Japanese people that showed opposite effects. Manual validation with the
relevant tweets showed an agreement with the statistical results. A case study
with Japanese tweets using supervised logistic regression classifies tweets
into heath-worry, economy-worry and other classes with 83.11% accuracy.
Predicted tweets around events re-confirm the statistical outcomes.
| 2,021 |
Computation and Language
|
Out-of-Scope Domain and Intent Classification through Hierarchical Joint
Modeling
|
User queries for a real-world dialog system may sometimes fall outside the
scope of the system's capabilities, but appropriate system responses will
enable smooth processing throughout the human-computer interaction. This paper
is concerned with the user's intent, and focuses on out-of-scope intent
classification in dialog systems. Although user intents are highly correlated
with the application domain, few studies have exploited such correlations for
intent classification. Rather than developing a two-stage approach that first
classifies the domain and then the intent, we propose a hierarchical multi-task
learning approach based on a joint model to classify domain and intent
simultaneously. Novelties in the proposed approach include: (1) sharing
supervised out-of-scope signals in joint modeling of domain and intent
classification to replace a two-stage pipeline; and (2) introducing a
hierarchical model that learns the intent and domain representations in the
higher and lower layers respectively. Experiments show that the model
outperforms existing methods in terms of accuracy, out-of-scope recall and F1.
Additionally, threshold-based post-processing further improves performance by
balancing precision and recall in intent classification.
| 2,021 |
Computation and Language
|
Mitigating Political Bias in Language Models Through Reinforced
Calibration
|
Current large-scale language models can be politically biased as a result of
the data they are trained on, potentially causing serious problems when they
are deployed in real-world settings. In this paper, we describe metrics for
measuring political bias in GPT-2 generation and propose a reinforcement
learning (RL) framework for mitigating political biases in generated text. By
using rewards from word embeddings or a classifier, our RL framework guides
debiased generation without having access to the training data or requiring the
model to be retrained. In empirical experiments on three attributes sensitive
to political bias (gender, location, and topic), our methods reduced bias
according to both our metrics and human evaluation, while maintaining
readability and semantic coherence.
| 2,021 |
Computation and Language
|
Scaling End-to-End Models for Large-Scale Multilingual ASR
|
Building ASR models across many languages is a challenging multi-task
learning problem due to large variations and heavily unbalanced data. Existing
work has shown positive transfer from high resource to low resource languages.
However, degradations on high resource languages are commonly observed due to
interference from the heterogeneous multilingual data and reduction in
per-language capacity. We conduct a capacity study on a 15-language task, with
the amount of data per language varying from 7.6K to 53.5K hours. We adopt
GShard [1] to efficiently scale up to 10B parameters. Empirically, we find that
(1) scaling the number of model parameters is an effective way to solve the
capacity bottleneck - our 500M-param model already outperforms monolingual
baselines and scaling it to 1B and 10B brought further quality gains; (2)
larger models are not only more data efficient, but also more efficient in
terms of training cost as measured in TPU days - the 1B-param model reaches the
same accuracy at 34% of training time as the 500M-param model; (3) given a
fixed capacity budget, adding depth works better than width and large encoders
do better than large decoders; (4) with continuous training, they can be
adapted to new languages and domains.
| 2,021 |
Computation and Language
|
The Factual Inconsistency Problem in Abstractive Text Summarization: A
Survey
|
Recently, various neural encoder-decoder models pioneered by Seq2Seq
framework have been proposed to achieve the goal of generating more abstractive
summaries by learning to map input text to output text. At a high level, such
neural models can freely generate summaries without any constraint on the words
or phrases used. Moreover, their format is closer to human-edited summaries and
output is more readable and fluent. However, the neural model's abstraction
ability is a double-edged sword. A commonly observed problem with the generated
summaries is the distortion or fabrication of factual information in the
article. This inconsistency between the original text and the summary has
caused various concerns over its applicability, and the previous evaluation
methods of text summarization are not suitable for this issue. In response to
the above problems, the current research direction is predominantly divided
into two categories, one is to design fact-aware evaluation metrics to select
outputs without factual inconsistency errors, and the other is to develop new
summarization systems towards factual consistency. In this survey, we focus on
presenting a comprehensive review of these fact-specific evaluation methods and
text summarization models.
| 2,023 |
Computation and Language
|
Summarization, Simplification, and Generation: The Case of Patents
|
We survey Natural Language Processing (NLP) approaches to summarizing,
simplifying, and generating patents' text. While solving these tasks has
important practical applications - given patents' centrality in the R&D process
- patents' idiosyncrasies open peculiar challenges to the current NLP state of
the art. This survey aims at a) describing patents' characteristics and the
questions they raise to the current NLP systems, b) critically presenting
previous work and its evolution, and c) drawing attention to directions of
research in which further work is needed. To the best of our knowledge, this is
the first survey of generative approaches in the patent domain.
| 2,022 |
Computation and Language
|
BERT Meets Relational DB: Contextual Representations of Relational
Databases
|
In this paper, we address the problem of learning low dimension
representation of entities on relational databases consisting of multiple
tables. Embeddings help to capture semantics encoded in the database and can be
used in a variety of settings like auto-completion of tables, fully-neural
query processing of relational joins queries, seamlessly handling missing
values, and more. Current work is restricted to working with just single table,
or using pretrained embeddings over an external corpus making them unsuitable
for use in real-world databases. In this work, we look into ways of using these
attention-based model to learn embeddings for entities in the relational
database. We are inspired by BERT style pretraining methods and are interested
in observing how they can be extended for representation learning on structured
databases. We evaluate our approach of the autocompletion of relational
databases and achieve improvement over standard baselines.
| 2,021 |
Computation and Language
|
Word-Level Alignment of Paper Documents with their Electronic Full-Text
Counterparts
|
We describe a simple procedure for the automatic creation of word-level
alignments between printed documents and their respective full-text versions.
The procedure is unsupervised, uses standard, off-the-shelf components only,
and reaches an F-score of 85.01 in the basic setup and up to 86.63 when using
pre- and post-processing. Potential areas of application are manual database
curation (incl. document triage) and biomedical expression OCR.
| 2,021 |
Computation and Language
|
GTN-ED: Event Detection Using Graph Transformer Networks
|
Recent works show that the graph structure of sentences, generated from
dependency parsers, has potential for improving event detection. However, they
often only leverage the edges (dependencies) between words, and discard the
dependency labels (e.g., nominal-subject), treating the underlying graph edges
as homogeneous. In this work, we propose a novel framework for incorporating
both dependencies and their labels using a recently proposed technique called
Graph Transformer Networks (GTN). We integrate GTNs to leverage dependency
relations on two existing homogeneous-graph-based models, and demonstrate an
improvement in the F1 score on the ACE dataset.
| 2,021 |
Computation and Language
|
Paraphrastic Representations at Scale
|
We present a system that allows users to train their own state-of-the-art
paraphrastic sentence representations in a variety of languages. We also
release trained models for English, Arabic, German, French, Spanish, Russian,
Turkish, and Chinese. We train these models on large amounts of data, achieving
significantly improved performance from the original papers proposing the
methods on a suite of monolingual semantic similarity, cross-lingual semantic
similarity, and bitext mining tasks. Moreover, the resulting models surpass all
prior work on unsupervised semantic textual similarity, significantly
outperforming even BERT-based models like Sentence-BERT (Reimers and Gurevych,
2019). Additionally, our models are orders of magnitude faster than prior work
and can be used on CPU with little difference in inference speed (even improved
speed over GPU when using more CPU cores), making these models an attractive
choice for users without access to GPUs or for use on embedded devices.
Finally, we add significantly increased functionality to the code bases for
training paraphrastic sentence models, easing their use for both inference and
for training them for any desired language with parallel data. We also include
code to automatically download and preprocess training data.
| 2,023 |
Computation and Language
|
Explanation-Based Human Debugging of NLP Models: A Survey
|
Debugging a machine learning model is hard since the bug usually involves the
training data and the learning process. This becomes even harder for an opaque
deep learning model if we have no clue about how the model actually works. In
this survey, we review papers that exploit explanations to enable humans to
give feedback and debug NLP models. We call this problem explanation-based
human debugging (EBHD). In particular, we categorize and discuss existing work
along three dimensions of EBHD (the bug context, the workflow, and the
experimental setting), compile findings on how EBHD components affect the
feedback providers, and highlight open problems that could be future research
directions.
| 2,021 |
Computation and Language
|
Leveraging Machine Learning to Detect Data Curation Activities
|
This paper describes a machine learning approach for annotating and analyzing
data curation work logs at ICPSR, a large social sciences data archive. The
systems we studied track curation work and coordinate team decision-making at
ICPSR. Repository staff use these systems to organize, prioritize, and document
curation work done on datasets, making them promising resources for studying
curation work and its impact on data reuse, especially in combination with data
usage analytics. A key challenge, however, is classifying similar activities so
that they can be measured and associated with impact metrics. This paper
contributes: 1) a schema of data curation activities; 2) a computational model
for identifying curation actions in work log descriptions; and 3) an analysis
of frequent data curation activities at ICPSR over time. We first propose a
schema of data curation actions to help us analyze the impact of curation work.
We then use this schema to annotate a set of data curation logs, which contain
records of data transformations and project management decisions completed by
repository staff. Finally, we train a text classifier to detect the frequency
of curation actions in a large set of work logs. Our approach supports the
analysis of curation work documented in work log systems as an important step
toward studying the relationship between research data curation and data reuse.
| 2,022 |
Computation and Language
|
An analysis of full-size Russian complexly NER labelled corpus of
Internet user reviews on the drugs based on deep learning and language neural
nets
|
We present the full-size Russian complexly NER-labeled corpus of Internet
user reviews, along with an evaluation of accuracy levels reached on this
corpus by a set of advanced deep learning neural networks to extract the
pharmacologically meaningful entities from Russian texts. The corpus annotation
includes mentions of the following entities: Medication (33005 mentions),
Adverse Drug Reaction (1778), Disease (17403), and Note (4490). Two of them -
Medication and Disease - comprise a set of attributes. A part of the corpus has
the coreference annotation with 1560 coreference chains in 300 documents.
Special multi-label model based on a language model and the set of features is
developed, appropriate for presented corpus labeling. The influence of the
choice of different modifications of the models: word vector representations,
types of language models pre-trained for Russian, text normalization styles,
and other preliminary processing are analyzed. The sufficient size of our
corpus allows to study the effects of particularities of corpus labeling and
balancing entities in the corpus. As a result, the state of the art for the
pharmacological entity extraction problem for Russian is established on a
full-size labeled corpus. In case of the adverse drug reaction (ADR)
recognition, it is 61.1 by the F1-exact metric that, as our analysis shows, is
on par with the accuracy level for other language corpora with similar
characteristics and the ADR representativnes. The evaluated baseline precision
of coreference relation extraction on the corpus is 71, that is higher the
results reached on other Russian corpora.
| 2,021 |
Computation and Language
|
Evaluating Attribution in Dialogue Systems: The BEGIN Benchmark
|
Knowledge-grounded dialogue systems powered by large language models often
generate responses that, while fluent, are not attributable to a relevant
source of information. Progress towards models that do not exhibit this issue
requires evaluation metrics that can quantify its prevalence. To this end, we
introduce the Benchmark for Evaluation of Grounded INteraction (BEGIN),
comprised of 12k dialogue turns generated by neural dialogue systems trained on
three knowledge-grounded dialogue corpora. We collect human annotations
assessing the extent to which the models' responses can be attributed to the
given background information. We then use BEGIN to analyze eight evaluation
metrics. We find that these metrics rely on spurious correlations, do not
reliably distinguish attributable abstractive responses from unattributable
ones, and perform substantially worse when the knowledge source is longer. Our
findings underscore the need for more sophisticated and robust evaluation
metrics for knowledge-grounded dialogue. We make BEGIN publicly available at
https://github.com/google/BEGIN-dataset.
| 2,022 |
Computation and Language
|
Improving Response Quality with Backward Reasoning in Open-domain
Dialogue Systems
|
Being able to generate informative and coherent dialogue responses is crucial
when designing human-like open-domain dialogue systems. Encoder-decoder-based
dialogue models tend to produce generic and dull responses during the decoding
step because the most predictable response is likely to be a non-informative
response instead of the most suitable one. To alleviate this problem, we
propose to train the generation model in a bidirectional manner by adding a
backward reasoning step to the vanilla encoder-decoder training. The proposed
backward reasoning step pushes the model to produce more informative and
coherent content because the forward generation step's output is used to infer
the dialogue context in the backward direction. The advantage of our method is
that the forward generation and backward reasoning steps are trained
simultaneously through the use of a latent variable to facilitate bidirectional
optimization. Our method can improve response quality without introducing side
information (e.g., a pre-trained topic model). The proposed bidirectional
response generation method achieves state-of-the-art performance for response
quality.
| 2,021 |
Computation and Language
|
Capturing Logical Structure of Visually Structured Documents with
Multimodal Transition Parser
|
While many NLP pipelines assume raw, clean texts, many texts we encounter in
the wild, including a vast majority of legal documents, are not so clean, with
many of them being visually structured documents (VSDs) such as PDFs.
Conventional preprocessing tools for VSDs mainly focused on word segmentation
and coarse layout analysis, whereas fine-grained logical structure analysis
(such as identifying paragraph boundaries and their hierarchies) of VSDs is
underexplored. To that end, we proposed to formulate the task as prediction of
"transition labels" between text fragments that maps the fragments to a tree,
and developed a feature-based machine learning system that fuses visual,
textual and semantic cues.Our system is easily customizable to different types
of VSDs and it significantly outperformed baselines in identifying different
structures in VSDs. For example, our system obtained a paragraph boundary
detection F1 score of 0.953 which is significantly better than a popular
PDF-to-text tool with an F1 score of 0.739.
| 2,021 |
Computation and Language
|
Hidden Backdoors in Human-Centric Language Models
|
Natural language processing (NLP) systems have been proven to be vulnerable
to backdoor attacks, whereby hidden features (backdoors) are trained into a
language model and may only be activated by specific inputs (called triggers),
to trick the model into producing unexpected behaviors. In this paper, we
create covert and natural triggers for textual backdoor attacks, \textit{hidden
backdoors}, where triggers can fool both modern language models and human
inspection. We deploy our hidden backdoors through two state-of-the-art trigger
embedding methods. The first approach via homograph replacement, embeds the
trigger into deep neural networks through the visual spoofing of lookalike
character replacement. The second approach uses subtle differences between text
generated by language models and real natural text to produce trigger sentences
with correct grammar and high fluency. We demonstrate that the proposed hidden
backdoors can be effective across three downstream security-critical NLP tasks,
representative of modern human-centric NLP systems, including toxic comment
detection, neural machine translation (NMT), and question answering (QA). Our
two hidden backdoor attacks can achieve an Attack Success Rate (ASR) of at
least $97\%$ with an injection rate of only $3\%$ in toxic comment detection,
$95.1\%$ ASR in NMT with less than $0.5\%$ injected data, and finally $91.12\%$
ASR against QA updated with only 27 poisoning data samples on a model
previously trained with 92,024 samples (0.029\%). We are able to demonstrate
the adversary's high success rate of attacks, while maintaining functionality
for regular users, with triggers inconspicuous by the human administrators.
| 2,021 |
Computation and Language
|
AlloST: Low-resource Speech Translation without Source Transcription
|
The end-to-end architecture has made promising progress in speech translation
(ST). However, the ST task is still challenging under low-resource conditions.
Most ST models have shown unsatisfactory results, especially in the absence of
word information from the source speech utterance. In this study, we survey
methods to improve ST performance without using source transcription, and
propose a learning framework that utilizes a language-independent universal
phone recognizer. The framework is based on an attention-based
sequence-to-sequence model, where the encoder generates the phonetic embeddings
and phone-aware acoustic representations, and the decoder controls the fusion
of the two embedding streams to produce the target token sequence. In addition
to investigating different fusion strategies, we explore the specific usage of
byte pair encoding (BPE), which compresses a phone sequence into a
syllable-like segmented sequence. Due to the conversion of symbols, a segmented
sequence represents not only pronunciation but also language-dependent
information lacking in phones. Experiments conducted on the Fisher
Spanish-English and Taigi-Mandarin drama corpora show that our method
outperforms the conformer-based baseline, and the performance is close to that
of the existing best method using source transcription.
| 2,022 |
Computation and Language
|
MRCBert: A Machine Reading ComprehensionApproach for Unsupervised
Summarization
|
When making an online purchase, it becomes important for the customer to read
the product reviews carefully and make a decision based on that. However,
reviews can be lengthy, may contain repeated, or sometimes irrelevant
information that does not help in decision making. In this paper, we introduce
MRCBert, a novel unsupervised method to generate summaries from product
reviews. We leverage Machine Reading Comprehension, i.e. MRC, approach to
extract relevant opinions and generate both rating-wise and aspect-wise
summaries from reviews. Through MRCBert we show that we can obtain reasonable
performance using existing models and transfer learning, which can be useful
for learning under limited or low resource scenarios. We demonstrated our
results on reviews of a product from the Electronics category in the Amazon
Reviews dataset. Our approach is unsupervised as it does not require any
domain-specific dataset, such as the product review dataset, for training or
fine-tuning. Instead, we have used SQuAD v1.1 dataset only to fine-tune BERT
for the MRC task. Since MRCBert does not require a task-specific dataset, it
can be easily adapted and used in other domains.
| 2,021 |
Computation and Language
|
It's not what you said, it's how you said it: discriminative perception
of speech as a multichannel communication system
|
People convey information extremely effectively through spoken interaction
using multiple channels of information transmission: the lexical channel of
what is said, and the non-lexical channel of how it is said. We propose
studying human perception of spoken communication as a means to better
understand how information is encoded across these channels, focusing on the
question 'What characteristics of communicative context affect listener's
expectations of speech?'. To investigate this, we present a novel behavioural
task testing whether listeners can discriminate between the true utterance in a
dialogue and utterances sampled from other contexts with the same lexical
content. We characterize how perception - and subsequent discriminative
capability - is affected by different degrees of additional contextual
information across both the lexical and non-lexical channel of speech. Results
demonstrate that people can effectively discriminate between different prosodic
realisations, that non-lexical context is informative, and that this channel
provides more salient information than the lexical channel, highlighting the
importance of the non-lexical channel in spoken interaction.
| 2,021 |
Computation and Language
|
PREDICT: Persian Reverse Dictionary
|
Finding the appropriate words to convey concepts (i.e., lexical access) is
essential for effective communication. Reverse dictionaries fulfill this need
by helping individuals to find the word(s) which could relate to a specific
concept or idea. To the best of our knowledge, this resource has not been
available for the Persian language. In this paper, we compare four different
architectures for implementing a Persian reverse dictionary (PREDICT).
We evaluate our models using (phrase,word) tuples extracted from the only
Persian dictionaries available online, namely Amid, Moein, and Dehkhoda where
the phrase describes the word. Given the phrase, a model suggests the most
relevant word(s) in terms of the ability to convey the concept. The model is
considered to perform well if the correct word is one of its top suggestions.
Our experiments show that a model consisting of Long Short-Term Memory (LSTM)
units enhanced by an additive attention mechanism is enough to produce
suggestions comparable to (or in some cases better than) the word in the
original dictionary. The study also reveals that the model sometimes produces
the synonyms of the word as its output which led us to introduce a new metric
for the evaluation of reverse dictionaries called Synonym Accuracy accounting
for the percentage of times the event of producing the word or a synonym of it
occurs. The assessment of the best model using this new metric also indicates
that at least 62% of the times, it produces an accurate result within the top
100 suggestions.
| 2,021 |
Computation and Language
|
When to Fold'em: How to answer Unanswerable questions
|
We present 3 different question-answering models trained on the SQuAD2.0
dataset -- BIDAF, DocumentQA and ALBERT Retro-Reader -- demonstrating the
improvement of language models in the past three years. Through our research in
fine-tuning pre-trained models for question-answering, we developed a novel
approach capable of achieving a 2% point improvement in SQuAD2.0 F1 in reduced
training time. Our method of re-initializing select layers of a
parameter-shared language model is simple yet empirically powerful.
| 2,021 |
Computation and Language
|
MathBERT: A Pre-Trained Model for Mathematical Formula Understanding
|
Large-scale pre-trained models like BERT, have obtained a great success in
various Natural Language Processing (NLP) tasks, while it is still a challenge
to adapt them to the math-related tasks. Current pre-trained models neglect the
structural features and the semantic correspondence between formula and its
context. To address these issues, we propose a novel pre-trained model, namely
\textbf{MathBERT}, which is jointly trained with mathematical formulas and
their corresponding contexts. In addition, in order to further capture the
semantic-level structural features of formulas, a new pre-training task is
designed to predict the masked formula substructures extracted from the
Operator Tree (OPT), which is the semantic structural representation of
formulas. We conduct various experiments on three downstream tasks to evaluate
the performance of MathBERT, including mathematical information retrieval,
formula topic classification and formula headline generation. Experimental
results demonstrate that MathBERT significantly outperforms existing methods on
all those three tasks. Moreover, we qualitatively show that this pre-trained
model effectively captures the semantic-level structural information of
formulas. To the best of our knowledge, MathBERT is the first pre-trained model
for mathematical formula understanding.
| 2,021 |
Computation and Language
|
Intelligent Conversational Android ERICA Applied to Attentive Listening
and Job Interview
|
Following the success of spoken dialogue systems (SDS) in smartphone
assistants and smart speakers, a number of communicative robots are developed
and commercialized. Compared with the conventional SDSs designed as a
human-machine interface, interaction with robots is expected to be in a closer
manner to talking to a human because of the anthropomorphism and physical
presence. The goal or task of dialogue may not be information retrieval, but
the conversation itself. In order to realize human-level "long and deep"
conversation, we have developed an intelligent conversational android ERICA. We
set up several social interaction tasks for ERICA, including attentive
listening, job interview, and speed dating. To allow for spontaneous,
incremental multiple utterances, a robust turn-taking model is implemented
based on TRP (transition-relevance place) prediction, and a variety of
backchannels are generated based on time frame-wise prediction instead of
IPU-based prediction. We have realized an open-domain attentive listening
system with partial repeats and elaborating questions on focus words as well as
assessment responses. It has been evaluated with 40 senior people, engaged in
conversation of 5-7 minutes without a conversation breakdown. It was also
compared against the WOZ setting. We have also realized a job interview system
with a set of base questions followed by dynamic generation of elaborating
questions. It has also been evaluated with student subjects, showing promising
results.
| 2,021 |
Computation and Language
|
Event Argument Extraction using Causal Knowledge Structures
|
Event Argument extraction refers to the task of extracting structured
information from unstructured text for a particular event of interest. The
existing works exhibit poor capabilities to extract causal event arguments like
Reason and After Effects. Furthermore, most of the existing works model this
task at a sentence level, restricting the context to a local scope. While it
may be effective for short spans of text, for longer bodies of text such as
news articles, it has often been observed that the arguments for an event do
not necessarily occur in the same sentence as that containing an event trigger.
To tackle the issue of argument scattering across sentences, the use of global
context becomes imperative in this task. In our work, we propose an external
knowledge aided approach to infuse document-level event information to aid the
extraction of complex event arguments. We develop a causal network for our
event-annotated dataset by extracting relevant event causal structures from
ConceptNet and phrases from Wikipedia. We use the extracted event causal
features in a bi-directional transformer encoder to effectively capture
long-range inter-sentence dependencies. We report the effectiveness of our
proposed approach through both qualitative and quantitative analysis. In this
task, we establish our findings on an event annotated dataset in 5 Indian
languages. This dataset adds further complexity to the task by labelling
arguments of entity type (like Time, Place) as well as more complex argument
types (like Reason, After-Effect). Our approach achieves state-of-the-art
performance across all the five languages. Since our work does not rely on any
language-specific features, it can be easily extended to other languages.
| 2,021 |
Computation and Language
|
Larger-Scale Transformers for Multilingual Masked Language Modeling
|
Recent work has demonstrated the effectiveness of cross-lingual language
model pretraining for cross-lingual understanding. In this study, we present
the results of two larger multilingual masked language models, with 3.5B and
10.7B parameters. Our two new models dubbed XLM-R XL and XLM-R XXL outperform
XLM-R by 1.8% and 2.4% average accuracy on XNLI. Our model also outperforms the
RoBERTa-Large model on several English tasks of the GLUE benchmark by 0.3% on
average while handling 99 more languages. This suggests pretrained models with
larger capacity may obtain both strong performance on high-resource languages
while greatly improving low-resource languages. We make our code and models
publicly available.
| 2,021 |
Computation and Language
|
Searchable Hidden Intermediates for End-to-End Models of Decomposable
Sequence Tasks
|
End-to-end approaches for sequence tasks are becoming increasingly popular.
Yet for complex sequence tasks, like speech translation, systems that cascade
several models trained on sub-tasks have shown to be superior, suggesting that
the compositionality of cascaded systems simplifies learning and enables
sophisticated search capabilities. In this work, we present an end-to-end
framework that exploits compositionality to learn searchable hidden
representations at intermediate stages of a sequence model using decomposed
sub-tasks. These hidden intermediates can be improved using beam search to
enhance the overall performance and can also incorporate external models at
intermediate stages of the network to re-score or adapt towards out-of-domain
data. One instance of the proposed framework is a Multi-Decoder model for
speech translation that extracts the searchable hidden intermediates from a
speech recognition sub-task. The model demonstrates the aforementioned benefits
and outperforms the previous state-of-the-art by around +6 and +3 BLEU on the
two test sets of Fisher-CallHome and by around +3 and +4 BLEU on the
English-German and English-French test sets of MuST-C.
| 2,021 |
Computation and Language
|
Amharic Text Clustering Using Encyclopedic Knowledge with Neural Word
Embedding
|
In this digital era, almost in every discipline people are using automated
systems that generate information represented in document format in different
natural languages. As a result, there is a growing interest towards better
solutions for finding, organizing and analyzing these documents. In this paper,
we propose a system that clusters Amharic text documents using Encyclopedic
Knowledge (EK) with neural word embedding. EK enables the representation of
related concepts and neural word embedding allows us to handle the contexts of
the relatedness. During the clustering process, all the text documents pass
through preprocessing stages. Enriched text document features are extracted
from each document by mapping with EK and word embedding model. TF-IDF weighted
vector of enriched feature was generated. Finally, text documents are clustered
using popular spherical K-means algorithm. The proposed system is tested with
Amharic text corpus and Amharic Wikipedia data. Test results show that the use
of EK with word embedding for document clustering improves the average accuracy
over the use of only EK. Furthermore, changing the size of the class has a
significant effect on accuracy.
| 2,022 |
Computation and Language
|
NaijaNER : Comprehensive Named Entity Recognition for 5 Nigerian
Languages
|
Most of the common applications of Named Entity Recognition (NER) is on
English and other highly available languages. In this work, we present our
findings on Named Entity Recognition for 5 Nigerian Languages (Nigerian
English, Nigerian Pidgin English, Igbo, Yoruba and Hausa). These languages are
considered low-resourced, and very little openly available Natural Language
Processing work has been done in most of them. In this work, individual NER
models were trained and metrics recorded for each of the languages. We also
worked on a combined model that can handle Named Entity Recognition (NER) for
any of the five languages. The combined model works well for Named Entity
Recognition(NER) on each of the languages and with better performance compared
to individual NER models trained specifically on annotated data for the
specific language. The aim of this work is to share our learning on how
information extraction using Named Entity Recognition can be optimized for the
listed Nigerian Languages for inclusion, ease of deployment in production and
reusability of models. Models developed during this project are available on
GitHub https://git.io/JY0kk and an interactive web app
https://nigner.herokuapp.com/.
| 2,021 |
Computation and Language
|
CBench: Towards Better Evaluation of Question Answering Over Knowledge
Graphs
|
Recently, there has been an increase in the number of knowledge graphs that
can be only queried by experts. However, describing questions using structured
queries is not straightforward for non-expert users who need to have sufficient
knowledge about both the vocabulary and the structure of the queried knowledge
graph, as well as the syntax of the structured query language used to describe
the user's information needs. The most popular approach introduced to overcome
the aforementioned challenges is to use natural language to query these
knowledge graphs. Although several question answering benchmarks can be used to
evaluate question-answering systems over a number of popular knowledge graphs,
choosing a benchmark to accurately assess the quality of a question answering
system is a challenging task.
In this paper, we introduce CBench, an extensible, and more informative
benchmarking suite for analyzing benchmarks and evaluating question answering
systems. CBench can be used to analyze existing benchmarks with respect to
several fine-grained linguistic, syntactic, and structural properties of the
questions and queries in the benchmark. We show that existing benchmarks vary
significantly with respect to these properties deeming choosing a small subset
of them unreliable in evaluating QA systems. Until further research improves
the quality and comprehensiveness of benchmarks, CBench can be used to
facilitate this evaluation using a set of popular benchmarks that can be
augmented with other user-provided benchmarks. CBench not only evaluates a
question answering system based on popular single-number metrics but also gives
a detailed analysis of the linguistic, syntactic, and structural properties of
answered and unanswered questions to better help the developers of question
answering systems to better understand where their system excels and where it
struggles.
| 2,021 |
Computation and Language
|
Layer Reduction: Accelerating Conformer-Based Self-Supervised Model via
Layer Consistency
|
Transformer-based self-supervised models are trained as feature extractors
and have empowered many downstream speech tasks to achieve state-of-the-art
performance. However, both the training and inference process of these models
may encounter prohibitively high computational cost and large parameter budget.
Although Parameter Sharing Strategy (PSS) proposed in ALBERT paves the way for
parameter reduction, the computation required remains the same. Interestingly,
we found in experiments that distributions of feature embeddings from different
Transformer layers are similar when PSS is integrated: a property termed as
Layer Consistency (LC) in this paper. Given this similarity of feature
distributions, we assume that feature embeddings from different layers would
have similar representing power. In this work, Layer Consistency enables us to
adopt Transformer-based models in a more efficient manner: the number of
Conformer layers in each training iteration could be uniformly sampled and
Shallow Layer Inference (SLI) could be applied to reduce the number of layers
in inference stage. In experiments, our models are trained with LibriSpeech
dataset and then evaluated on both phone classification and Speech Recognition
tasks. We experimentally achieve 7.8X parameter reduction, 41.9% training
speedup and 37.7% inference speedup while maintaining comparable performance
with conventional BERT-like self-supervised methods.
| 2,021 |
Computation and Language
|
Transformers: "The End of History" for NLP?
|
Recent advances in neural architectures, such as the Transformer, coupled
with the emergence of large-scale pre-trained models such as BERT, have
revolutionized the field of Natural Language Processing (NLP), pushing the
state of the art for a number of NLP tasks. A rich family of variations of
these models has been proposed, such as RoBERTa, ALBERT, and XLNet, but
fundamentally, they all remain limited in their ability to model certain kinds
of information, and they cannot cope with certain information sources, which
was easy for pre-existing models. Thus, here we aim to shed light on some
important theoretical limitations of pre-trained BERT-style models that are
inherent in the general Transformer architecture. First, we demonstrate in
practice on two general types of tasks -- segmentation and segment labeling --
and on four datasets that these limitations are indeed harmful and that
addressing them, even in some very simple and naive ways, can yield sizable
improvements over vanilla RoBERTa and XLNet models. Then, we offer a more
general discussion on desiderata for future additions to the Transformer
architecture that would increase its expressiveness, which we hope could help
in the design of the next generation of deep NLP architectures.
| 2,021 |
Computation and Language
|
Representation Learning for Weakly Supervised Relation Extraction
|
Recent years have seen rapid development in Information Extraction, as well
as its subtask, Relation Extraction. Relation Extraction is able to detect
semantic relations between entities in sentences. Currently, many efficient
approaches have been applied to relation extraction tasks. Supervised learning
approaches especially have good performance. However, there are still many
difficult challenges. One of the most serious problems is that manually labeled
data is difficult to acquire. In most cases, limited data for supervised
approaches equals lousy performance. Thus here, under the situation with only
limited training data, we focus on how to improve the performance of our
supervised baseline system with unsupervised pre-training. Feature is one of
the key components in improving the supervised approaches. Traditional
approaches usually apply hand-crafted features, which require expert knowledge
and expensive human labor. However, this type of feature might suffer from data
sparsity: when the training set size is small, the model parameters might be
poorly estimated. In this thesis, we present several novel unsupervised
pre-training models to learn the distributed text representation features,
which are encoded with rich syntactic-semantic patterns of relation
expressions. The experiments have demonstrated that this type of feature,
combine with the traditional hand-crafted features, could improve the
performance of the logistic classification model for relation extraction,
especially on the classification of relations with only minor training
instances.
| 2,024 |
Computation and Language
|
What's in a Summary? Laying the Groundwork for Advances in
Hospital-Course Summarization
|
Summarization of clinical narratives is a long-standing research problem.
Here, we introduce the task of hospital-course summarization. Given the
documentation authored throughout a patient's hospitalization, generate a
paragraph that tells the story of the patient admission. We construct an
English, text-to-text dataset of 109,000 hospitalizations (2M source notes) and
their corresponding summary proxy: the clinician-authored "Brief Hospital
Course" paragraph written as part of a discharge note. Exploratory analyses
reveal that the BHC paragraphs are highly abstractive with some long extracted
fragments; are concise yet comprehensive; differ in style and content
organization from the source notes; exhibit minimal lexical cohesion; and
represent silver-standard references. Our analysis identifies multiple
implications for modeling this complex, multi-document summarization task.
| 2,021 |
Computation and Language
|
BERT based freedom to operate patent analysis
|
In this paper we present a method to apply BERT to freedom to operate patent
analysis and patent searches. According to the method, BERT is fine-tuned by
training patent descriptions to the independent claims. Each description
represents an invention which is protected by the corresponding claims. Such a
trained BERT could be able to identify or order freedom to operate relevant
patents based on a short description of an invention or product. We tested the
method by training BERT on the patent class G06T1/00 and applied the trained
BERT on five inventions classified in G06T1/60, described via DOCDB abstracts.
The DOCDB abstract are available on ESPACENET of the European Patent Office.
| 2,021 |
Computation and Language
|
Measuring diachronic sense change: new models and Monte Carlo methods
for Bayesian inference
|
In a bag-of-words model, the senses of a word with multiple meanings, e.g.
"bank" (used either in a river-bank or an institution sense), are represented
as probability distributions over context words, and sense prevalence is
represented as a probability distribution over senses. Both of these may change
with time. Modelling and measuring this kind of sense change is challenging due
to the typically high-dimensional parameter space and sparse datasets. A
recently published corpus of ancient Greek texts contains expert-annotated
sense labels for selected target words. Automatic sense-annotation for the word
"kosmos" (meaning decoration, order or world) has been used as a test case in
recent work with related generative models and Monte Carlo methods. We adapt an
existing generative sense change model to develop a simpler model for the main
effects of sense and time, and give MCMC methods for Bayesian inference on all
these models that are more efficient than existing methods. We carry out
automatic sense-annotation of snippets containing "kosmos" using our model, and
measure the time-evolution of its three senses and their prevalence. As far as
we are aware, ours is the first analysis of this data, within the class of
generative models we consider, that quantifies uncertainty and returns credible
sets for evolving sense prevalence in good agreement with those given by expert
annotation.
| 2,022 |
Computation and Language
|
Switching Contexts: Transportability Measures for NLP
|
This paper explores the topic of transportability, as a sub-area of
generalisability. By proposing the utilisation of metrics based on
well-established statistics, we are able to estimate the change in performance
of NLP models in new contexts. Defining a new measure for transportability may
allow for better estimation of NLP system performance in new domains, and is
crucial when assessing the performance of NLP systems in new tasks and domains.
Through several instances of increasing complexity, we demonstrate how
lightweight domain similarity measures can be used as estimators for the
transportability in NLP applications. The proposed transportability measures
are evaluated in the context of Named Entity Recognition and Natural Language
Inference tasks.
| 2,021 |
Computation and Language
|
A Survey of Recent Abstract Summarization Techniques
|
This paper surveys several recent abstract summarization methods: T5,
Pegasus, and ProphetNet. We implement the systems in two languages: English and
Indonesian languages. We investigate the impact of pre-training models (one T5,
three Pegasuses, three ProphetNets) on several Wikipedia datasets in English
and Indonesian language and compare the results to the Wikipedia systems'
summaries. The T5-Large, the Pegasus-XSum, and the ProphetNet-CNNDM provide the
best summarization. The most significant factors that influence ROUGE
performance are coverage, density, and compression. The higher the scores, the
better the summary. Other factors that influence the ROUGE scores are the
pre-training goal, the dataset's characteristics, the dataset used for testing
the pre-trained model, and the cross-lingual function. Several suggestions to
improve this paper's limitation are: 1) assure that the dataset used for the
pre-training model must sufficiently large, contains adequate instances for
handling cross-lingual purpose; 2) Advanced process (finetuning) shall be
reasonable. We recommend using the large dataset consists of comprehensive
coverage of topics from many languages before implementing advanced processes
such as the train-infer-train procedure to the zero-shot translation in the
training stage of the pre-training model.
| 2,021 |
Computation and Language
|
DEUX: An Attribute-Guided Framework for Sociable Recommendation Dialog
Systems
|
It is important for sociable recommendation dialog systems to perform as both
on-task content and social content to engage users and gain their favor. In
addition to understand the user preferences and provide a satisfying
recommendation, such systems must be able to generate coherent and natural
social conversations to the user. Traditional dialog state tracking cannot be
applied to such systems because it does not track the attributes in the social
content. To address this challenge, we propose DEUX, a novel attribute-guided
framework to create better user experiences while accomplishing a movie
recommendation task. DEUX has a module that keeps track of the movie attributes
(e.g., favorite genres, actors,etc.) in both user utterances and system
responses. This allows the system to introduce new movie attributes in its
social content. Then, DEUX has multiple values for the same attribute type
which suits the recommendation task since a user may like multiple genres, for
instance. Experiments suggest that DEUX outperforms all the baselines on being
more consistent, fitting the user preferences better, and providing a more
engaging chat experience. Our approach can be used for any similar problems of
sociable task-oriented dialog system.
| 2,021 |
Computation and Language
|
WhatTheWikiFact: Fact-Checking Claims Against Wikipedia
|
The rise of Internet has made it a major source of information.
Unfortunately, not all information online is true, and thus a number of
fact-checking initiatives have been launched, both manual and automatic, to
deal with the problem. Here, we present our contribution in this regard:
\emph{WhatTheWikiFact}, a system for automatic claim verification using
Wikipedia. The system can predict the veracity of an input claim, and it
further shows the evidence it has retrieved as part of the verification
process. It shows confidence scores and a list of relevant Wikipedia articles,
together with detailed information about each article, including the phrase
used to retrieve it, the most relevant sentences extracted from it and their
stance with respect to the input claim, as well as the associated
probabilities. The system supports several languages: Bulgarian, English, and
Russian.
| 2,021 |
Computation and Language
|
AMMU : A Survey of Transformer-based Biomedical Pretrained Language
Models
|
Transformer-based pretrained language models (PLMs) have started a new era in
modern natural language processing (NLP). These models combine the power of
transformers, transfer learning, and self-supervised learning (SSL). Following
the success of these models in the general domain, the biomedical research
community has developed various in-domain PLMs starting from BioBERT to the
latest BioELECTRA and BioALBERT models. We strongly believe there is a need for
a survey paper that can provide a comprehensive survey of various
transformer-based biomedical pretrained language models (BPLMs). In this
survey, we start with a brief overview of foundational concepts like
self-supervised learning, embedding layer and transformer encoder layers. We
discuss core concepts of transformer-based PLMs like pretraining methods,
pretraining tasks, fine-tuning methods, and various embedding types specific to
biomedical domain. We introduce a taxonomy for transformer-based BPLMs and then
discuss all the models. We discuss various challenges and present possible
solutions. We conclude by highlighting some of the open issues which will drive
the research community to further improve transformer-based BPLMs.
| 2,021 |
Computation and Language
|
Memorisation versus Generalisation in Pre-trained Language Models
|
State-of-the-art pre-trained language models have been shown to memorise
facts and perform well with limited amounts of training data. To gain a better
understanding of how these models learn, we study their generalisation and
memorisation capabilities in noisy and low-resource scenarios. We find that the
training of these models is almost unaffected by label noise and that it is
possible to reach near-optimal results even on extremely noisy datasets.
However, our experiments also show that they mainly learn from high-frequency
patterns and largely fail when tested on low-resource tasks such as few-shot
learning and rare entity recognition. To mitigate such limitations, we propose
an extension based on prototypical networks that improves performance in
low-resource named entity recognition tasks.
| 2,022 |
Computation and Language
|
Natural Language Generation Using Link Grammar for General
Conversational Intelligence
|
Many current artificial general intelligence (AGI) and natural language
processing (NLP) architectures do not possess general conversational
intelligence--that is, they either do not deal with language or are unable to
convey knowledge in a form similar to the human language without manual,
labor-intensive methods such as template-based customization. In this paper, we
propose a new technique to automatically generate grammatically valid sentences
using the Link Grammar database. This natural language generation method far
outperforms current state-of-the-art baselines and may serve as the final
component in a proto-AGI question answering pipeline that understandably
handles natural language material.
| 2,021 |
Computation and Language
|
Federated Word2Vec: Leveraging Federated Learning to Encourage
Collaborative Representation Learning
|
Large scale contextual representation models have significantly advanced NLP
in recent years, understanding the semantics of text to a degree never seen
before. However, they need to process large amounts of data to achieve
high-quality results. Joining and accessing all these data from multiple
sources can be extremely challenging due to privacy and regulatory reasons.
Federated Learning can solve these limitations by training models in a
distributed fashion, taking advantage of the hardware of the devices that
generate the data. We show the viability of training NLP models, specifically
Word2Vec, with the Federated Learning protocol. In particular, we focus on a
scenario in which a small number of organizations each hold a relatively large
corpus. The results show that neither the quality of the results nor the
convergence time in Federated Word2Vec deteriorates as compared to centralised
Word2Vec.
| 2,021 |
Computation and Language
|
Semantic Journeys: Quantifying Change in Emoji Meaning from 2012-2018
|
The semantics of emoji has, to date, been considered from a static
perspective. We offer the first longitudinal study of how emoji semantics
changes over time, applying techniques from computational linguistics to six
years of Twitter data. We identify five patterns in emoji semantic development
and find evidence that the less abstract an emoji is, the more likely it is to
undergo semantic change. In addition, we analyse select emoji in more detail,
examining the effect of seasonality and world events on emoji semantics. To aid
future work on emoji and semantics, we make our data publicly available along
with a web-based interface that anyone can use to explore semantic change in
emoji.
| 2,021 |
Computation and Language
|
Teaching NLP outside Linguistics and Computer Science classrooms: Some
challenges and some opportunities
|
NLP's sphere of influence went much beyond computer science research and the
development of software applications in the past decade. We see people using
NLP methods in a range of academic disciplines from Asian Studies to Clinical
Oncology. We also notice the presence of NLP as a module in most of the data
science curricula within and outside of regular university setups. These
courses are taken by students from very diverse backgrounds. This paper takes a
closer look at some issues related to teaching NLP to these diverse audiences
based on my classroom experiences, and identifies some challenges the
instructors face, particularly when there is no ecosystem of related courses
for the students. In this process, it also identifies a few challenge areas for
both NLP researchers and tool developers.
| 2,021 |
Computation and Language
|
Pseudo Siamese Network for Few-shot Intent Generation
|
Few-shot intent detection is a challenging task due to the scare annotation
problem. In this paper, we propose a Pseudo Siamese Network (PSN) to generate
labeled data for few-shot intents and alleviate this problem. PSN consists of
two identical subnetworks with the same structure but different weights: an
action network and an object network. Each subnetwork is a transformer-based
variational autoencoder that tries to model the latent distribution of
different components in the sentence. The action network is learned to
understand action tokens and the object network focuses on object-related
expressions. It provides an interpretable framework for generating an utterance
with an action and an object existing in a given intent. Experiments on two
real-world datasets show that PSN achieves state-of-the-art performance for the
generalized few shot intent detection task.
| 2,021 |
Computation and Language
|
Impact of Gender Debiased Word Embeddings in Language Modeling
|
Gender, race and social biases have recently been detected as evident
examples of unfairness in applications of Natural Language Processing. A key
path towards fairness is to understand, analyse and interpret our data and
algorithms. Recent studies have shown that the human-generated data used in
training is an apparent factor of getting biases. In addition, current
algorithms have also been proven to amplify biases from data.
To further address these concerns, in this paper, we study how an
state-of-the-art recurrent neural language model behaves when trained on data,
which under-represents females, using pre-trained standard and debiased word
embeddings. Results show that language models inherit higher bias when trained
on unbalanced data when using pre-trained embeddings, in comparison with using
embeddings trained within the task. Moreover, results show that, on the same
data, language models inherit lower bias when using debiased pre-trained
emdeddings, compared to using standard pre-trained embeddings.
| 2,021 |
Computation and Language
|
Trends, Limitations and Open Challenges in Automatic Readability
Assessment Research
|
Readability assessment is the task of evaluating the reading difficulty of a
given piece of text. Although research on computational approaches to
readability assessment is now two decades old, there is not much work on
synthesizing this research. This article is a brief survey of contemporary
research on developing computational models for readability assessment. We
identify the common approaches, discuss their shortcomings, and identify some
challenges for the future. Where possible, we also connect computational
research with insights from related work in other disciplines such as education
and psychology.
| 2,022 |
Computation and Language
|
Russian News Clustering and Headline Selection Shared Task
|
This paper presents the results of the Russian News Clustering and Headline
Selection shared task. As a part of it, we propose the tasks of Russian news
event detection, headline selection, and headline generation. These tasks are
accompanied by datasets and baselines. The presented datasets for event
detection and headline selection are the first public Russian datasets for
their tasks. The headline generation dataset is based on clustering and
provides multiple reference headlines for every cluster, unlike the previous
datasets. Finally, the approaches proposed by the shared task participants are
reported and analyzed.
| 2,021 |
Computation and Language
|
On the limit of English conversational speech recognition
|
In our previous work we demonstrated that a single headed attention
encoder-decoder model is able to reach state-of-the-art results in
conversational speech recognition. In this paper, we further improve the
results for both Switchboard 300 and 2000. Through use of an improved
optimizer, speaker vector embeddings, and alternative speech representations we
reduce the recognition errors of our LSTM system on Switchboard-300 by 4%
relative. Compensation of the decoder model with the probability ratio approach
allows more efficient integration of an external language model, and we report
5.9% and 11.5% WER on the SWB and CHM parts of Hub5'00 with very simple LSTM
models. Our study also considers the recently proposed conformer, and more
advanced self-attention based language models. Overall, the conformer shows
similar performance to the LSTM; nevertheless, their combination and decoding
with an improved LM reaches a new record on Switchboard-300, 5.0% and 10.0% WER
on SWB and CHM. Our findings are also confirmed on Switchboard-2000, and a new
state of the art is reported, practically reaching the limit of the benchmark.
| 2,021 |
Computation and Language
|
Leveraging Deep Representations of Radiology Reports in Survival
Analysis for Predicting Heart Failure Patient Mortality
|
Utilizing clinical texts in survival analysis is difficult because they are
largely unstructured. Current automatic extraction models fail to capture
textual information comprehensively since their labels are limited in scope.
Furthermore, they typically require a large amount of data and high-quality
expert annotations for training. In this work, we present a novel method of
using BERT-based hidden layer representations of clinical texts as covariates
for proportional hazards models to predict patient survival outcomes. We show
that hidden layers yield notably more accurate predictions than predefined
features, outperforming the previous baseline model by 5.7% on average across
C-index and time-dependent AUC. We make our work publicly available at
https://github.com/bionlplab/heart_failure_mortality.
| 2,021 |
Computation and Language
|
SUPERB: Speech processing Universal PERformance Benchmark
|
Self-supervised learning (SSL) has proven vital for advancing research in
natural language processing (NLP) and computer vision (CV). The paradigm
pretrains a shared model on large volumes of unlabeled data and achieves
state-of-the-art (SOTA) for various tasks with minimal adaptation. However, the
speech processing community lacks a similar setup to systematically explore the
paradigm. To bridge this gap, we introduce Speech processing Universal
PERformance Benchmark (SUPERB). SUPERB is a leaderboard to benchmark the
performance of a shared model across a wide range of speech processing tasks
with minimal architecture changes and labeled data. Among multiple usages of
the shared model, we especially focus on extracting the representation learned
from SSL due to its preferable re-usability. We present a simple framework to
solve SUPERB tasks by learning task-specialized lightweight prediction heads on
top of the frozen shared model. Our results demonstrate that the framework is
promising as SSL representations show competitive generalizability and
accessibility across SUPERB tasks. We release SUPERB as a challenge with a
leaderboard and a benchmark toolkit to fuel the research in representation
learning and general speech processing.
| 2,021 |
Computation and Language
|
Applied Language Technology: NLP for the Humanities
|
This contribution describes a two-course module that seeks to provide
humanities majors with a basic understanding of language technology and its
applications using Python. The learning materials consist of interactive
Jupyter Notebooks and accompanying YouTube videos, which are openly available
with a Creative Commons licence.
| 2,021 |
Computation and Language
|
Modeling Social Readers: Novel Tools for Addressing Reception from
Online Book Reviews
|
Readers' responses to literature have received scant attention in
computational literary studies. The rise of social media offers an opportunity
to capture a segment of these responses while data-driven analysis of these
responses can provide new critical insight into how people "read". Posts
discussing an individual book on Goodreads, a social media platform that hosts
user discussions of popular literature, are referred to as "reviews", and
consist of plot summaries, opinions, quotes, or some mixture of these. Since
these reviews are written by readers, computationally modeling them allows one
to discover the overall non-professional discussion space about a work,
including an aggregated summary of the work's plot, an implicit ranking of the
importance of events, and the readers' impressions of main characters. We
develop a pipeline of interlocking computational tools to extract a
representation of this reader generated shared narrative model. Using a corpus
of reviews of five popular novels, we discover the readers' distillation of the
main storylines in a novel, their understanding of the relative importance of
characters, as well as the readers' varying impressions of these characters. In
so doing, we make three important contributions to the study of infinite
vocabulary networks: (i) an automatically derived narrative network that
includes meta-actants; (ii) a new sequencing algorithm, REV2SEQ, that generates
a consensus sequence of events based on partial trajectories aggregated from
the reviews; and (iii) a new "impressions" algorithm, SENT2IMP, that provides
finer, non-trivial and multi-modal insight into readers' opinions of
characters.
| 2,021 |
Computation and Language
|
Scalar Adjective Identification and Multilingual Ranking
|
The intensity relationship that holds between scalar adjectives (e.g., nice <
great < wonderful) is highly relevant for natural language inference and
common-sense reasoning. Previous research on scalar adjective ranking has
focused on English, mainly due to the availability of datasets for evaluation.
We introduce a new multilingual dataset in order to promote research on scalar
adjectives in new languages. We perform a series of experiments and set
performance baselines on this dataset, using monolingual and multilingual
contextual language models. Additionally, we introduce a new binary
classification task for English scalar adjective identification which examines
the models' ability to distinguish scalar from relational adjectives. We probe
contextualised representations and report baseline results for future
comparison on this task.
| 2,021 |
Computation and Language
|
Unreasonable Effectiveness of Rule-Based Heuristics in Solving Russian
SuperGLUE Tasks
|
Leader-boards like SuperGLUE are seen as important incentives for active
development of NLP, since they provide standard benchmarks for fair comparison
of modern language models. They have driven the world's best engineering teams
as well as their resources to collaborate and solve a set of tasks for general
language understanding. Their performance scores are often claimed to be close
to or even higher than the human performance. These results encouraged more
thorough analysis of whether the benchmark datasets featured any statistical
cues that machine learning based language models can exploit. For English
datasets, it was shown that they often contain annotation artifacts. This
allows solving certain tasks with very simple rules and achieving competitive
rankings.
In this paper, a similar analysis was done for the Russian SuperGLUE (RSG), a
recently published benchmark set and leader-board for Russian natural language
understanding. We show that its test datasets are vulnerable to shallow
heuristics. Often approaches based on simple rules outperform or come close to
the results of the notorious pre-trained language models like GPT-3 or BERT. It
is likely (as the simplest explanation) that a significant part of the SOTA
models performance in the RSG leader-board is due to exploiting these shallow
heuristics and that has nothing in common with real language understanding. We
provide a set of recommendations on how to improve these datasets, making the
RSG leader-board even more representative of the real progress in Russian NLU.
| 2,021 |
Computation and Language
|
ZEN 2.0: Continue Training and Adaption for N-gram Enhanced Text
Encoders
|
Pre-trained text encoders have drawn sustaining attention in natural language
processing (NLP) and shown their capability in obtaining promising results in
different tasks. Recent studies illustrated that external self-supervised
signals (or knowledge extracted by unsupervised learning, such as n-grams) are
beneficial to provide useful semantic evidence for understanding languages such
as Chinese, so as to improve the performance on various downstream tasks
accordingly. To further enhance the encoders, in this paper, we propose to
pre-train n-gram-enhanced encoders with a large volume of data and advanced
techniques for training. Moreover, we try to extend the encoder to different
languages as well as different domains, where it is confirmed that the same
architecture is applicable to these varying circumstances and new
state-of-the-art performance is observed from a long list of NLP tasks across
languages and domains.
| 2,021 |
Computation and Language
|
Semantic Extractor-Paraphraser based Abstractive Summarization
|
The anthology of spoken languages today is inundated with textual
information, necessitating the development of automatic summarization models.
In this manuscript, we propose an extractor-paraphraser based abstractive
summarization system that exploits semantic overlap as opposed to its
predecessors that focus more on syntactic information overlap. Our model
outperforms the state-of-the-art baselines in terms of ROUGE, METEOR and word
mover similarity (WMS), establishing the superiority of the proposed system via
extensive ablation experiments. We have also challenged the summarization
capabilities of the state of the art Pointer Generator Network (PGN), and
through thorough experimentation, shown that PGN is more of a paraphraser,
contrary to the prevailing notion of a summarizer; illustrating it's
incapability to accumulate information across multiple sentences.
| 2,021 |
Computation and Language
|
Discourse Relation Embeddings: Representing the Relations between
Discourse Segments in Social Media
|
Discourse relations are typically modeled as a discrete class that
characterizes the relation between segments of text (e.g. causal explanations,
expansions). However, such predefined discrete classes limits the universe of
potential relationships and their nuanced differences. Analogous to contextual
word embeddings, we propose representing discourse relations as points in high
dimensional continuous space. However, unlike words, discourse relations often
have no surface form (relations are between two segments, often with no word or
phrase in that gap) which presents a challenge for existing embedding
techniques. We present a novel method for automatically creating discourse
relation embeddings (DiscRE), addressing the embedding challenge through a
weakly supervised, multitask approach to learn diverse and nuanced relations
between discourse segments in social media. Results show DiscRE can: (1) obtain
the best performance on Twitter discourse relation classification task (macro
F1=0.76) (2) improve the state of the art in social media causality prediction
(from F1=.79 to .81), (3) perform beyond modern sentence and contextual word
embeddings at traditional discourse relation classification, and (4) capture
novel nuanced relations (e.g. relations semantically at the intersection of
causal explanations and counterfactuals).
| 2,023 |
Computation and Language
|
Inferring the Reader: Guiding Automated Story Generation with
Commonsense Reasoning
|
Transformer-based language model approaches to automated story generation
currently provide state-of-the-art results. However, they still suffer from
plot incoherence when generating narratives over time, and critically lack
basic commonsense reasoning. Furthermore, existing methods generally focus only
on single-character stories, or fail to track characters at all. To improve the
coherence of generated narratives and to expand the scope of character-centric
narrative generation, we introduce Commonsense-inference Augmented neural
StoryTelling (CAST), a framework for introducing commonsense reasoning into the
generation process with the option to model the interaction between multiple
characters. We find that our CAST method produces significantly more coherent,
on-topic, enjoyable and fluent stories than existing models in both the
single-character and two-character settings in three storytelling domains.
| 2,023 |
Computation and Language
|
BLM-17m: A Large-Scale Dataset for Black Lives Matter Topic Detection on
Twitter
|
Protection of human rights is one of the most important problems of our
world. In this paper, our aim is to provide a dataset which covers one of the
most significant human rights contradiction in recent months affected the whole
world, George Floyd incident. We propose a labeled dataset for topic detection
that contains 17 million tweets. These Tweets are collected from 25 May 2020 to
21 August 2020 that covers 89 days from start of this incident. We labeled the
dataset by monitoring most trending news topics from global and local
newspapers. Apart from that, we present two baselines, TF-IDF and LDA. We
evaluated the results of these two methods with three different k values for
metrics of precision, recall and f1-score. The collected dataset is available
at https://github.com/MeysamAsgariC/BLMT.
| 2,023 |
Computation and Language
|
GraphTMT: Unsupervised Graph-based Topic Modeling from Video Transcripts
|
To unfold the tremendous amount of multimedia data uploaded daily to social
media platforms, effective topic modeling techniques are needed. Existing work
tends to apply topic models on written text datasets. In this paper, we propose
a topic extractor on video transcripts. Exploiting neural word embeddings
through graph-based clustering, we aim to improve usability and semantic
coherence. Unlike most topic models, this approach works without knowing the
true number of topics, which is important when no such assumption can or should
be made. Experimental results on the real-life multimodal dataset MuSe-CaR
demonstrates that our approach GraphTMT extracts coherent and meaningful topics
and outperforms baseline methods. Furthermore, we successfully demonstrate the
applicability of our approach on the popular Citysearch corpus.
| 2,021 |
Computation and Language
|
Conversational Machine Reading Comprehension for Vietnamese Healthcare
Texts
|
Machine reading comprehension (MRC) is a sub-field in natural language
processing that aims to assist computers understand unstructured texts and then
answer questions related to them. In practice, the conversation is an essential
way to communicate and transfer information. To help machines understand
conversation texts, we present UIT-ViCoQA, a new corpus for conversational
machine reading comprehension in the Vietnamese language. This corpus consists
of 10,000 questions with answers over 2,000 conversations about health news
articles. Then, we evaluate several baseline approaches for conversational
machine comprehension on the UIT-ViCoQA corpus. The best model obtains an F1
score of 45.27%, which is 30.91 points behind human performance (76.18%),
indicating that there is ample room for improvement. Our dataset is available
at our website: http://nlp.uit.edu.vn/datasets/ for research purposes.
| 2,021 |
Computation and Language
|
Data Augmentation by Concatenation for Low-Resource Translation: A
Mystery and a Solution
|
In this paper, we investigate the driving factors behind concatenation, a
simple but effective data augmentation method for low-resource neural machine
translation. Our experiments suggest that discourse context is unlikely the
cause for the improvement of about +1 BLEU across four language pairs. Instead,
we demonstrate that the improvement comes from three other factors unrelated to
discourse: context diversity, length diversity, and (to a lesser extent)
position shifting.
| 2,021 |
Computation and Language
|
HerBERT: Efficiently Pretrained Transformer-based Language Model for
Polish
|
BERT-based models are currently used for solving nearly all Natural Language
Processing (NLP) tasks and most often achieve state-of-the-art results.
Therefore, the NLP community conducts extensive research on understanding these
models, but above all on designing effective and efficient training procedures.
Several ablation studies investigating how to train BERT-like models have been
carried out, but the vast majority of them concerned only the English language.
A training procedure designed for English does not have to be universal and
applicable to other especially typologically different languages. Therefore,
this paper presents the first ablation study focused on Polish, which, unlike
the isolating English language, is a fusional language. We design and
thoroughly evaluate a pretraining procedure of transferring knowledge from
multilingual to monolingual BERT-based models. In addition to multilingual
model initialization, other factors that possibly influence pretraining are
also explored, i.e. training objective, corpus size, BPE-Dropout, and
pretraining length. Based on the proposed procedure, a Polish BERT-based
language model -- HerBERT -- is trained. This model achieves state-of-the-art
results on multiple downstream tasks.
| 2,021 |
Computation and Language
|
ExcavatorCovid: Extracting Events and Relations from Text Corpora for
Temporal and Causal Analysis for COVID-19
|
Timely responses from policy makers to mitigate the impact of the COVID-19
pandemic rely on a comprehensive grasp of events, their causes, and their
impacts. These events are reported at such a speed and scale as to be
overwhelming. In this paper, we present ExcavatorCovid, a machine reading
system that ingests open-source text documents (e.g., news and scientific
publications), extracts COVID19 related events and relations between them, and
builds a Temporal and Causal Analysis Graph (TCAG). Excavator will help
government agencies alleviate the information overload, understand likely
downstream effects of political and economic decisions and events related to
the pandemic, and respond in a timely manner to mitigate the impact of
COVID-19. We expect the utility of Excavator to outlive the COVID-19 pandemic:
analysts and decision makers will be empowered by Excavator to better
understand and solve complex problems in the future. An interactive TCAG
visualization is available at http://afrl402.bbn.com:5050/index.html. We also
released a demonstration video at https://vimeo.com/528619007.
| 2,021 |
Computation and Language
|
Full-Sentence Models Perform Better in Simultaneous Translation Using
the Information Enhanced Decoding Strategy
|
Simultaneous translation, which starts translating each sentence after
receiving only a few words in source sentence, has a vital role in many
scenarios. Although the previous prefix-to-prefix framework is considered
suitable for simultaneous translation and achieves good performance, it still
has two inevitable drawbacks: the high computational resource costs caused by
the need to train a separate model for each latency $k$ and the insufficient
ability to encode information because each target token can only attend to a
specific source prefix. We propose a novel framework that adopts a simple but
effective decoding strategy which is designed for full-sentence models. Within
this framework, training a single full-sentence model can achieve arbitrary
given latency and save computational resources. Besides, with the competence of
the full-sentence model to encode the whole sentence, our decoding strategy can
enhance the information maintained in the decoded states in real time.
Experimental results show that our method achieves better translation quality
than baselines on 4 directions: Zh$\rightarrow$En, En$\rightarrow$Ro and
En$\leftrightarrow$De.
| 2,022 |
Computation and Language
|
Mind Reading at Work: Cooperation without common ground
|
As Stefan Kopp and Nicole Kramer say in their recent paper[Frontiers in
Psychology 12 (2021) 597], despite some very impressive demonstrations over the
last decade or so, we still don't know how how to make a computer have a half
decent conversation with a human. They argue that the capabilities required to
do this include incremental joint co-construction and mentalizing. Although
agreeing whole heartedly with their statement of the problem, this paper argues
for a different approach to the solution based on the "new" AI of situated
action.
| 2,021 |
Computation and Language
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.